October 13, 2026
Rabbet's moat in real estate development software
SaaS is dead, so every SaaS company still standing today calls itself an AI company.
What does this mean exactly?
Well, we all know that AI writes code for “free,” and if code is free, then by extension so is adding a feature to your product. At the same time, AI has changed what’s possible to ship. The unautomatable is now table stakes—as we like to say: AI is alchemy.
Nonetheless, business problems remain unsolved. Your customers never really cared about the tech you used. They only care about the end result.
What have you done for them lately?
In 2023, it was possible that if a startup said “AI” in just the right tone of voice, its revenue accelerated. Sad to say, that moment is gone.
Neither SaaS nor AI are descriptions of the job to be done, which leaves customers right back where they always were: trying to figure out which vendor will actually do the job they need. This is the “state of nature” for customers: even when there are dozens of startups offering to help them with “AI transformation,” it’s hard to find one that can actually deliver.
And in finance, the job to be done will always require a level of solemn guarantees—the physics of finance—that are difficult if not impossible for AI to achieve. This is the singular challenge of any technology company in finance: you have to execute the job to be done and prove that it’s right every time.
AI in contact with atoms and dollars
Nowhere is that challenge more visible right now than in real estate development, an industry that has quietly drawn its own line about where AI belongs and where it doesn’t.
For Rabbet, a QED portfolio company, the job to be done is simple: developers need to tell their lenders exactly what has happened on a project in the last month in order to get lenders to give them money for the next month. Lenders need to analyze that information, make sure it’s complete and in line with expectations and then send the money out. Rabbet automates both.
In the status quo, these jobs today are done by humans. Someone managing six active development projects has to reconstruct the story of last month in Excel formulas, Outlook threads and shared folders that nobody names the same way. The emails contain invoices, our hero downloads it, reads it, decides where it fits and enters it by hand in Excel. Proof that the work was complete—called a lien waiver—turns up days later in a different inbox from a different sender. When all the work is done, tied up in another mammoth PDF called a draw package, it’s been a dozen hours across four systems that don’t speak, and the numbers in the package are already stale.
But the real killer isn’t this process of data entry isn’t actually the killer, it’s that all these documents, all this work, hasn’t answered the routine business questions that actually matter: Where are we against budget? How much equity is left? What’s overrunning?
Real estate development has already drawn the line
This sounds like an opportunity for AI. And it is. But real developers who are allocating money for real shovels and real concrete in real dirt don’t want “AI” yet. Developers trust AI to read. They don’t trust it to calculate. In Rabbet’s 2026 State of Construction Finance Report, 71 percent said they trust AI to read text, but only 21 percent trust it to handle calculations.
Developers aren’t scared of AI, but they are experts in risk management. Trust but verify. When you use AI to reading, verification is easy: AI pulls a vendor name or an invoice date, a human glances at it, and moves on. Calculations are hard, humans can’t do them in their heads; and the numbers really matter.
Moreover, calculations are high stakes. If you give a wrong number to your lender, the whole project could halt.
Where AI earns its place
The document layer is high-volume and arrives in no consistent format. Invoices, lien waivers and contracts come in constantly, with no shared template and nothing attached to say where they belong. Reading through that pile is slow work that no one wants and it’s work where a mistake surfaces immediately instead of compounding.
The calculation stays deterministic, which is what an audit asks for: show where the number came from and show that it hasn’t drifted.
Every tool a real estate developer can buy was built for a job. Procore was built to run the jobsite. Cost tools were built to report on the budget. Excel was built to calculate. None of those is the job of knowing where six projects stand and being able to prove it.
Spreadsheets and email are the real incumbent in this category. Excel is flexible, universally understood and free at the margin, which is why it survives well past the point where it should. What it can’t do is read a document, hold an audit trail or update itself. Every number is manual labor, and every reconciliation costs an afternoon.
Rabbet Development is real estate development management software for owners, developers and asset managers that creates a connected financial system across the project’s budget, anticipated costs, draw schedule and funding sources.
Its differentiation is not simply that it captures project data or forecasts costs—it continuously connects those data points to the financial mechanics of the deal. As commitments, invoices, change orders and draws occur, Rabbet updates the forward-looking cost and funding picture, allowing teams to see emerging exposure early, understand its impact on contingency and capital and act before a variance becomes a problem. In other words, Rabbet is designed around the financial control layer of development: every cost has a place in the budget, every budget change has a funding implication, and the project’s financial position stays current as reality changes.
Northspyre also uses document intelligence to capture and analyze project costs, so document reading itself is not a meaningful point of differentiation and is getting rapidly commoditized with new frontier models. The sharper distinction is what happens next. Northspyre is designed around broad development management, cost intelligence and analytics; focused on creating an intelligence layer and reporting on projects.
Procore is construction project management software built for field and project execution. Its center of gravity is the jobsite: scheduling, submittals, RFIs, coordination across trades. While it has financial modules, development teams typically run it alongside a financial system rather than in place of one.
Customers have a job to do
The developer’s job is knowing where six projects stand on a Tuesday and defending their numbers to a lender on Wednesday.
When developers submit documents to a lender, millions of dollars are on the line. AI can help you read the docs and enter the data, but you can’t have AI freelance on the calculations. That’s why the future of tech in finance must bridge a system of record and a system of intelligence.
Rabbet Development is built on that distinction: AI handles the paper, the system of record handles the money. Structure data at the front door, and project costs stop being a spreadsheet someone rebuilt last Tuesday, they become the benchmark for the next job and the broader portfolio. That compounding clarity is the real moat; the AI model is just a commodity.
**
About Rabbet
Rabbet is a QED-backed real estate software company with two products. Rabbet’s mission is to take the friction out of development execution through structured, connected data. Rabbet helps real estate teams turn fragmented project information into trusted insights that can be shared, understood, and acted on throughout the development lifecycle.
Rabbet Development is real estate development management software for owners, developers and asset managers. It brings budgets, anticipated costs, draw schedules, and funding sources together across every phase of development, so teams always know where every project stands, answer questions without chasing information, and spot financial changes before they become project problems.
Rabbet Construction Lending is construction loan management software for banks, credit unions and private lenders. It brings loan administration, documentation, draw review and portfolio reporting into one system, so lending policies are applied consistently, portfolio risk is visible as projects evolve and every funding decision is backed by the evidence required to move money.
SaaS is dead, so every SaaS company still standing today calls itself an AI company.
What does this mean exactly?
Well, we all know that AI writes code for “free,” and if code is free, then by extension so is adding a feature to your product. At the same time, AI has changed what’s possible to ship. The unautomatable is now table stakes—as we like to say: AI is alchemy.
Nonetheless, business problems remain unsolved. Your customers never really cared about the tech you used. They only care about the end result.
What have you done for them lately?
In 2023, it was possible that if a startup said “AI” in just the right tone of voice, its revenue accelerated. Sad to say, that moment is gone.
Neither SaaS nor AI are descriptions of the job to be done, which leaves customers right back where they always were: trying to figure out which vendor will actually do the job they need. This is the “state of nature” for customers: even when there are dozens of startups offering to help them with “AI transformation,” it’s hard to find one that can actually deliver.
And in finance, the job to be done will always require a level of solemn guarantees—the physics of finance—that are difficult if not impossible for AI to achieve. This is the singular challenge of any technology company in finance: you have to execute the job to be done and prove that it’s right every time.
AI in contact with atoms and dollars
Nowhere is that challenge more visible right now than in real estate development, an industry that has quietly drawn its own line about where AI belongs and where it doesn’t.
For Rabbet, a QED portfolio company, the job to be done is simple: developers need to tell their lenders exactly what has happened on a project in the last month in order to get lenders to give them money for the next month. Lenders need to analyze that information, make sure it’s complete and in line with expectations and then send the money out. Rabbet automates both.
In the status quo, these jobs today are done by humans. Someone managing six active development projects has to reconstruct the story of last month in Excel formulas, Outlook threads and shared folders that nobody names the same way. The emails contain invoices, our hero downloads it, reads it, decides where it fits and enters it by hand in Excel. Proof that the work was complete—called a lien waiver—turns up days later in a different inbox from a different sender. When all the work is done, tied up in another mammoth PDF called a draw package, it’s been a dozen hours across four systems that don’t speak, and the numbers in the package are already stale.
But the real killer isn’t this process of data entry isn’t actually the killer, it’s that all these documents, all this work, hasn’t answered the routine business questions that actually matter: Where are we against budget? How much equity is left? What’s overrunning?
Real estate development has already drawn the line
This sounds like an opportunity for AI. And it is. But real developers who are allocating money for real shovels and real concrete in real dirt don’t want “AI” yet. Developers trust AI to read. They don’t trust it to calculate. In Rabbet’s 2026 State of Construction Finance Report, 71 percent said they trust AI to read text, but only 21 percent trust it to handle calculations.
Developers aren’t scared of AI, but they are experts in risk management. Trust but verify. When you use AI to reading, verification is easy: AI pulls a vendor name or an invoice date, a human glances at it, and moves on. Calculations are hard, humans can’t do them in their heads; and the numbers really matter.
Moreover, calculations are high stakes. If you give a wrong number to your lender, the whole project could halt.
Where AI earns its place
The document layer is high-volume and arrives in no consistent format. Invoices, lien waivers and contracts come in constantly, with no shared template and nothing attached to say where they belong. Reading through that pile is slow work that no one wants and it’s work where a mistake surfaces immediately instead of compounding.
The calculation stays deterministic, which is what an audit asks for: show where the number came from and show that it hasn’t drifted.
Every tool a real estate developer can buy was built for a job. Procore was built to run the jobsite. Cost tools were built to report on the budget. Excel was built to calculate. None of those is the job of knowing where six projects stand and being able to prove it.
Spreadsheets and email are the real incumbent in this category. Excel is flexible, universally understood and free at the margin, which is why it survives well past the point where it should. What it can’t do is read a document, hold an audit trail or update itself. Every number is manual labor, and every reconciliation costs an afternoon.
Rabbet Development is real estate development management software for owners, developers and asset managers that creates a connected financial system across the project’s budget, anticipated costs, draw schedule and funding sources.
Its differentiation is not simply that it captures project data or forecasts costs—it continuously connects those data points to the financial mechanics of the deal. As commitments, invoices, change orders and draws occur, Rabbet updates the forward-looking cost and funding picture, allowing teams to see emerging exposure early, understand its impact on contingency and capital and act before a variance becomes a problem. In other words, Rabbet is designed around the financial control layer of development: every cost has a place in the budget, every budget change has a funding implication, and the project’s financial position stays current as reality changes.
Northspyre also uses document intelligence to capture and analyze project costs, so document reading itself is not a meaningful point of differentiation and is getting rapidly commoditized with new frontier models. The sharper distinction is what happens next. Northspyre is designed around broad development management, cost intelligence and analytics; focused on creating an intelligence layer and reporting on projects.
Procore is construction project management software built for field and project execution. Its center of gravity is the jobsite: scheduling, submittals, RFIs, coordination across trades. While it has financial modules, development teams typically run it alongside a financial system rather than in place of one.
Customers have a job to do
The developer’s job is knowing where six projects stand on a Tuesday and defending their numbers to a lender on Wednesday.
When developers submit documents to a lender, millions of dollars are on the line. AI can help you read the docs and enter the data, but you can’t have AI freelance on the calculations. That’s why the future of tech in finance must bridge a system of record and a system of intelligence.
Rabbet Development is built on that distinction: AI handles the paper, the system of record handles the money. Structure data at the front door, and project costs stop being a spreadsheet someone rebuilt last Tuesday, they become the benchmark for the next job and the broader portfolio. That compounding clarity is the real moat; the AI model is just a commodity.
**
About Rabbet
Rabbet is a QED-backed real estate software company with two products. Rabbet’s mission is to take the friction out of development execution through structured, connected data. Rabbet helps real estate teams turn fragmented project information into trusted insights that can be shared, understood, and acted on throughout the development lifecycle.
Rabbet Development is real estate development management software for owners, developers and asset managers. It brings budgets, anticipated costs, draw schedules, and funding sources together across every phase of development, so teams always know where every project stands, answer questions without chasing information, and spot financial changes before they become project problems.
Rabbet Construction Lending is construction loan management software for banks, credit unions and private lenders. It brings loan administration, documentation, draw review and portfolio reporting into one system, so lending policies are applied consistently, portfolio risk is visible as projects evolve and every funding decision is backed by the evidence required to move money.